Outcomes of COVID-19 related hospitalisation among people with HIV in the ISARIC WHO Clinical Characterisation Protocol UK Protocol: prospective observational study Authors Anna Maria Geretti* 1 Alexander J. Stockdale* 1 Sophie H. Kelly* 1 Muge Cevik 2 Simon Collins 3 Laura Waters 4,5 Giovanni Villa 6 Annemarie Docherty 7,8 Ewen M Harrison 7 Lance Turtle 1 Peter JM Openshaw 9 J Kenneth Baillie 8,10 Caroline A. Sabin 11,12 $ Malcolm G Semple 1,13 $ *Joint first authors $ Joint senior authors CHASE study group** **Daniel Bradshaw 14 , Alison Brown 14 , Nicky Connor 14 , Valerie Delpech 14 , Saye Khoo 1 , Tamyo Mbisa 12, 14 , Chloe Orkin 15 , Ann Sullivan 16 . ISARIC4C Investigators [Listed separately] Affiliations 1. National Institute for Health Research (NIHR) Health Protection Research Unit (HPRU) in Emerging and Zoonotic Infections, Institute of Infection, Veterinary and Ecological Sciences, Faculty of Health and Life Sciences, University of Liverpool, Liverpool, UK 2. Division of Infection and Global Health Research, School of Medicine, University of St Andrews, St Andrews, UK All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449 doi: medRxiv preprint NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice.
32
Embed
Outcomes of COVID-19 related hospitalisation among people ......2020/08/07 · Outcomes of COVID-19 related hospitalisation among people with HIV in the ISARIC WHO Clinical Characterisation
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
Outcomes of COVID-19 related hospitalisation among people with HIV in the ISARIC 1
WHO Clinical Characterisation Protocol UK Protocol: prospective observational study 2
Tamyo Mbisa12, 14, Chloe Orkin15, Ann Sullivan16. 25
26
ISARIC4C Investigators [Listed separately] 27
28
Affiliations 29
1. National Institute for Health Research (NIHR) Health Protection Research Unit (HPRU) 30
in Emerging and Zoonotic Infections, Institute of Infection, Veterinary and Ecological 31
Sciences, Faculty of Health and Life Sciences, University of Liverpool, Liverpool, UK 32
2. Division of Infection and Global Health Research, School of Medicine, University of St 33
Andrews, St Andrews, UK 34
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice.
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
AMG - designed the study concept, reviewed all aspects of the data analysis and 67
interpretation, contributed to the writing of the manuscript, and performed the final review of 68
the manuscript. 69
AJS - performed the data analysis and contributed to the data interpretation and the writing of 70
the manuscript. 71
SHK - performed the initial literature search and contributed to the data analysis and 72
interpretation and the writing of the manuscript. 73
MC - contributed to conceptualisation and review of the manuscript 74
LW - contributed to conceptualisation and review of the manuscript 75
SC - contributed to conceptualisation and review of the manuscript 76
GV - contributed to data analysis and interpretation and the writing of the manuscript 77
AD - contributed to the ISARIC CCP-UK study design and data collection and reviewed the 78
manuscript 79
EMH - contributed to the ISARIC-CCP UK study design and data collection and reviewed 80
the manuscript 81
LT- contributed to the ISARIC-CCP UK study design and data collection and reviewed the 82
manuscript 83
PJMO - ISARIC CCP-UK Co-Lead investigator, sourced funding, contributed to the ISARIC 84
CCP-UK study design and data collection and reviewed the manuscript 85
JKB - ISARIC CCP-UK Consortium lead investigator, sourced funding, contributed to the 86
ISARIC CCP-UK study design and data collection and reviewed the manuscript 87
CAS - advised on all aspects of the conceptualisation and data analysis and interpretation, 88
and contributed to the writing and final review of the manuscript 89
MGS - ISARIC CCP-UK Protocol Chief Investigator and guarantor of the data, sourced 90
permissions and funding, contributed to the ISARIC CCP-UK study design and data 91
collection and reviewed the manuscript. 92
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
p=0·06). Following additional adjustment for disease severity at presentation, mortality was 113
higher in HIV-positive people (adjusted HR 1·63; 95% CI 1·07-2·48; p=0·02). In the HIV-114
positive group, mortality was more common among those who were slightly older and among 115
people with obesity and diabetes with complications. 116
Interpretation. HIV-positive status may be associated with an increased risk of day-28 117
mortality following a COVID-19 related hospitalisation. 118
Funding. NIHR, MRC, Wellcome Trust, Department for International Development, Bill and 119
Melinda Gates Foundation. 120
Study registration ISRCTN66726260 121
122
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
disease, and malignancy. Unlike some of the other evidence to date, but in line with the data 151
from South Africa, this study indicates that HIV-positive status may increase the risk of 152
mortality with COVID-19 compared to the general population, with an effect that was 153
especially evident among people with HIV aged below 60 years and was independent of 154
gender or ethnicity. Although we detected an association between mortality among people 155
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
with HIV and occurrence of obesity and diabetes with complication, the effect of HIV-156
positive status persisted after adjusting for comorbidities. 157
158
Implications of all the available evidence 159
People with HIV may be at increased risk of severe outcomes from COVID-19 compared to 160
the general population. Ongoing data collection is needed to confirm this association. 161
Linkage of hospital outcome data to the HIV history will be paramount to establishing the 162
determinants of the increased risk. COVID-19 related hospitalisation should pursue 163
systematic recording of HIV status to ensure optimal management and gathering of evidence. 164
165
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
obstructive pulmonary disease (COPD), serious cardiovascular disease, type II diabetes, and 169
transplant-related immunosuppression.1-5 There is no conclusive evidence about the 170
relationship with HIV infection. If untreated, HIV causes progressive immunosuppression; 171
however, antiretroviral therapy (ART) restores immune function and life-expectancy.6 172
Immune restoration is not always complete despite effective ART, however, and a subset of 173
people with HIV (PWH) remains at risk of persistent immune dysfunction,7 which might 174
augment severity of COVID-19, or conversely, possibly reduce the immune responses that 175
can complicate COVID-19.8 Although some antiretroviral drugs have been proposed to 176
protect against COVID-19, the evidence remains uncertain.9,10 Importantly, HIV might 177
increase the risk of adverse COVID-19 outcomes due to the common prevalence of co-factors 178
such as CKD, COPD, and diabetes,11 alongside socioeconomic variables that may carry a 179
negative influence.12 180
181
Several case series and observational cohort studies have described the outcomes of COVID-182
19 in PWH across Europe,9,13-19 Asia,18,19 and the United States.8,18-22 These studies have 183
often been limited by small sample size, lack of direct comparative data from people without 184
HIV, or inability to adjust for comorbidities. Some reports from Italy and New York 185
indicated that HIV did not increase the risk of COVID-19 related hospitalisation or 186
mortality,4,14,21 whereas two others suggested an increased risk of mortality among PWH 187
hospitalised with COVID-19.9,20 Preliminary data from South Africa similarly suggest that 188
HIV-positive status more than doubled the risk of COVID-19 related mortality.23 189
190
To characterise the presenting characteristics and outcomes of COVID-19 related 191
hospitalisation in PWH relative to those without HIV in the United Kingdom (UK), we 192
analysed data collected within the International Severe Acute Respiratory and emerging 193
Infections Consortium (ISARIC) Clinical Characterisation Protocol (CCP), the largest 194
prospective observational study of patients admitted to hospital with COVID-19 worldwide.24 195
196
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
detection in nasopharyngeal swabs was the only test available during the study and the 206
decision to test was at the discretion of the attending clinical team, who also decided upon 207
hospital admission, transfer into critical care and use of ventilation. For the present analyses, 208
baseline was defined as the date of hospital admission or symptom onset (for those with 209
symptom onset after hospitalisation, see below). Our analyses included individuals with a 210
baseline date that was on or before 4th June 2020 for whom �14 days had elapsed until the 211
date of data extraction on 18th June 2020. Individuals without information on the date of 212
admission or with a baseline date after 4th June 2020 were excluded. Where the date of 213
symptom onset was missing, we assumed that symptoms began on the date of the SARS-214
COV-2 PCR test, or if this was not recorded, the date of admission. Information on positive 215
HIV status, as reported to ISARIC CCP-UK, was confirmed through cross-checking with 216
reported receipt of ART (n=103), receipt of Pneumocystis jirovecii prophylaxis in the 217
absence of non-HIV indications (n=2), or directly with a site investigator (n=10). Individuals 218
with missing HIV status and those with unconfirmed HIV-positive status were excluded from 219
the analyses. 220
221
Statistical analysis 222
Presenting characteristics were compared between HIV-positive and HIV-negative people 223
and between PWH who died and those who survived to discharge using Wilcoxon rank sum 224
tests (for continuous variables) and Pearson's chi-squared or Fisher’s exact test (for 225
categorical variables). For all individuals, follow-up ended on the date of death. Patients 226
discharged to receive palliative care in the community were considered to have died three 227
days following discharge. Follow-up was right censored at day 28 for those remaining alive 228
as an inpatient, or for those who were discharged not for palliative care prior to day 28. A 229
data check showed that the vast majority of those discharged prior to day 28 were still alive 230
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
malignancy, and chronic haematological disease). We also included adjustment for the 243
baseline date to account for changes in mortality over the period of interest. Where entries on 244
comorbidity (presence or absence) were partially missing from the study CRF, we assumed 245
that missing data indicated the absence of the specific comorbidity; however, participants 246
with missing entries on all comorbidities were excluded from these adjusted analyses. 247
Finally, we fitted a further model with additional adjustment for hypoxia at presentation, 248
defined as oxygen saturation (SpO2) <94% on air or a record of receiving oxygen, as a 249
marker of presenting disease severity, in order to assess whether any increased/decreased risk 250
of mortality in PWH could be explained by a different stage of disease advancement at 251
hospitalisation. A series of sensitivity analyses were performed for the main mortality 252
outcome: i) we repeated the analyses after censoring follow-up on the day of discharge for 253
those discharged before day 28; ii) we included those with definite hospital-acquired COVID-254
19 (baseline = date of symptom onset); iii) we used symptom onset date as the baseline date 255
for all (rather than admission date where applicable); iv) we excluded PWH lacking a record 256
of ART; v) we calculated propensity scores for HIV-positive status using a logistic regression 257
model based on sex, ethnicity, age (in quadratic form), indeterminate/probable hospital 258
acquisition of COVID-19, smoking status, baseline date, and ten comorbidities, and included 259
the propensity score in a Cox regression model for death at 28 days; and vi) we considered a 260
binary endpoint of 14-day mortality and performed logistic regression (with the same 261
confounder adjustment as described above). In the HIV-positive group, we used a Cox 262
proportional hazard model to investigate the associations of presenting characteristics with 263
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
day-28 mortality. Analyses were conducted in Stata v16·1 (Statcorp, College Station, TX, 264
USA). 265
266
267
Results 268
Participants 269
ISARIC CCP-UK recorded 53,992 people with COVID-19 between 17th January 2020 and 270
18th June 2020. After excluding non-eligible participants (Figure 1), the final analysis 271
included 47,539 patients, of whom 115 (0·24%) had confirmed HIV-positive status. The 272
characteristics of patients excluded from the analysis did not differ by sex, ethnicity or age; in 273
particular, the characteristics of those excluded due to missing data on HIV status closely 274
resembled those reported to be HIV-negative (Supplementary Table 1). Among PWH, one 275
person was diagnosed with HIV during the admission and 103 (89·6%) had an ART record. 276
The regional distribution of study participants with HIV compared to the total UK population 277
of people accessing HIV care (2018 data) is shown in Supplementary Table 2. 278
279
Characteristics at presentation 280
The presenting characteristics according to HIV status are summarized in Tables 1-3 and 281
Figure 2. PWH were younger than HIV-negative people (medians of 55 versus 74 years, 282
p<0·001) (Table 1, Figure 2). There were fewer women in the HIV-positive group but 283
significantly larger proportions of people of black ethnicity. A similar proportion had no 284
recorded comorbidities, whereas occurrence of �2 comorbidities was more prevalent in the 285
HIV-negative group. PWH had lower prevalence of chronic cardiac disease, chronic 286
pulmonary disease, chronic neurological disorders, dementia, malignancy and 287
rheumatological disease, and higher rates of obesity and moderate/severe liver disease. There 288
were small differences in the prevalence of asthma, diabetes without complications, mild 289
liver disease, and malnutrition, whereas proportions with CKD, diabetes with complications 290
and chronic haematological disease were similar in the two groups. 291
292
The duration of symptoms was longer in the HIV-positive group (medians of 5 vs. 3 days, 293
p=0·001) (Table 2). PWH were more likely to present with systemic symptoms and signs, 294
including fever, headache, myalgia and tachycardia, and to have cough, sore throat and chest 295
pain. To a lesser extent, they also had more common occurrence of gastrointestinal 296
symptoms. Respiratory rate, occurrence of tachypnoea and hypoxia, and radiological 297
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
of COVID-19 and ten co-morbidities, the risk of mortality was 49% higher in PWH (adjusted 328
HR 1·49; 95% CI 0·99-2·26; p=0·06). Following additional adjustment for disease severity at 329
presentation (based on a record of hypoxia or oxygen therapy), the risk of mortality was 63% 330
higher in PWH (adjusted HR 1·63; 95% CI 1·07- 2·48; p=0·02) (Table 4). After day 28, there 331
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
censoring follow-up on the day of discharge for those discharged before day 28, including 336
patients with definite hospital acquired COVID-19, using symptom onset as the start of 337
follow-up, or excluding PWH lacking an ART record did not significantly alter the model. A 338
separate logistic regression model with a binary variable of day-14 mortality showed 339
increased odds of mortality in the HIV-positive group (adjusted OR 1·77; 95% CI 1·06-2·95; 340
p=0·03). 341
342
In the HIV-positive group, relative to patients who survived by day 28, patients who died 343
were slightly older and had a higher prevalence of obesity and diabetes with complications 344
(Table 5 and Supplementary Tables 6 and 7). An ART record was more often missing among 345
those who died. No indications of other major differences were observed. 346
347
Discussion 348
Principal findings 349
In this study of 115 HIV-positive and 47,979 HIV-negative people, we found evidence 350
suggesting a 63% increased risk of day-28 mortality among PWH hospitalised with COVID-351
19 compared to HIV-negative individuals in the same dataset, after adjustment for sex, 352
ethnicity, age, baseline date, ten key comorbidities, and disease severity at presentation (as 353
indicated by a record of hypoxia or receiving oxygen therapy). The latter adjustment took 354
into consideration that doctors may be more likely to admit HIV-positive people with 355
COVID-19 despite less severe symptoms. A striking difference in mortality was seen in the 356
younger age groups (<50 years and 50-59 years), although the number of older PWH was 357
small. 358
359
The role of age, sex and ethnicity on COVID-19 outcomes is the focus of much research.1,2,25 360
PWH in our study were significantly younger than the HIV-negative group and adjusting for 361
age changed the direction of the association between HIV status and day-28 mortality, 362
suggesting that age was a significant confounder in our analyses. Men were more prevalent in 363
the HIV-positive group, which is consistent with the epidemiology of HIV infection in the 364
UK, where men represent just over two thirds of the whole population with HIV.26 People of 365
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
black ethnicity made up nearly 42% of the HIV-positive group, whereas men and women of 366
black African ethnicity account for ~26% of the total number of PWH in the UK.26 367
Nonetheless, adjustment for sex or ethnicity alone did not impact our relative hazard 368
estimates. 369
370
Whilst there is a recognised interplay between HIV and comorbidities, omitting the 371
adjustment for comorbidities did not modify the association. PWH had fewer comorbidities, 372
notably lower prevalence of chronic pulmonary disease and malignancies, and this is likely to 373
be partly a function of their younger age. HIV-positive people who died were older and were 374
more likely to suffer from obesity and diabetes with complications than those who survived 375
to discharge. Similar trends have been seen in the general population.1,25 While these 376
observations highlight the importance of obesity and diabetes as cofactors, adjustment for 377
comorbidities in our model did not modify the association, suggesting that the apparent 378
increased risk of COVID-19 related mortality in PWH was not merely due to the presence of 379
promoting comorbidities. 380
381
Comparison with other studies 382
Evidence from published studies is not entirely consistent about the interplay between HIV 383
and COVID-19.8,9,13-23 A case-control study from New York compared 88 PWH, all of whom 384
were receiving ART, and 405 HIV-negative controls matched by age, gender, ethnicity, and 385
calendar week of infection.21 The study found no difference in the outcomes of COVID-19 386
related hospitalisation after adjusting for demographics, COPD, smoking, and baseline 387
ferritin and white blood cell count. Apart from the differences in the study design and 388
geography, there are fundamental differences between our study population and the New 389
York cohort described by Sigel et al.21 Most importantly, participants in the latter study were 390
older, with a median age of 61 years (IQR 54-67 years), whereas we found that the excess 391
mortality occurred in HIV-positive people aged below 60 years. Whereas malignancies were 392
recorded less commonly (3% vs. 10%), prevalence of obesity was nearly double in our cohort 393
(18% vs. 11%). At the other spectrum, preliminary data from the Western Cape Department 394
of Health in South Africa indicate that HIV-positive status was associated with increased 395
hazard of mortality [adjusted HR 2·75].23 Although the South African model did not account 396
for history of tuberculosis, obesity and socioeconomic status, it is of significance that HIV 397
suppression on ART made no difference to the risk. 398
399
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
A key strength of our study is the ability to perform a direct comparison of people with and 401
without HIV in the same dataset. Our analysis does not address risk factors for a COVID-19 402
diagnosis or a COVID-19 related hospitalisation among PWH, and cannot add to the current 403
debate about the role of certain antiretroviral agents in modulating such risks.9,10 In addition, 404
due to the format of data collection in ISARIC CCP-UK, our analysis cannot provide 405
evidence of the role of HIV-related parameters on outcomes of COVID-19 related 406
hospitalisation, as we did not have details of the ART history, current and nadir CD4 cell 407
count, plasma HIV-1 RNA load, and history of previous HIV-related disease. Only a subset 408
of CRFs from participants with HIV included a record of receiving ART and the records were 409
frequently incomplete. 410
411
Numerically, HIV-positive people who died were more likely not to have a record of being 412
on ART than those who remained alive at day 28. However, it cannot be stated with certainty 413
that those lacking an ART record were untreated nor therefore that lack of ART played a role 414
in the adverse outcomes. Our experience of working with large HIV datasets is that we often 415
find that people with missing data have worse mortality outcomes, simply because mortality 416
prevents collection of a detailed treatment history. In the UK, 93% of the 103,000 people 417
estimated to have HIV infection have been diagnosed and of these the vast majority (97%) 418
receives ART and maintains excellent suppression of the infection.26 Among those with a 419
HIV diagnosis, only a small subset of ~3% is either not engaged with care or experiences 420
problems with virological control despite ART.26 This suggests that the likelihood of PWH in 421
our study being off ART despite an absent record was overall low. 422
423
It is currently unclear how HIV infection and associated immune dysfunction modulates 424
infection with SARS CoV-2. Whilst immunosuppression was associated with poor COVID-425
19 outcomes in a recent meta-analysis,5 effective ART leads to immune reconstitution with 426
improved or normalised CD4 cell counts.6 We found no evidence of increased lymphopenia 427
among PWH in our study. Furthermore, compared to HIV-negative people, PWH were more 428
likely to experience systemic symptoms with fever and also showed higher CRP levels. These 429
observations are likely to be reflective of the younger age of PWH in our study,27 and at the 430
same time indicate preserved inflammatory responses in this group. This suggests that the 431
likelihood of PWH in our study being severely immunosuppressed was overall low. 432
433
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
After careful considerations and multiple adjustments for demographics, comorbidities and 435
disease severity on admission, our initial analyses of the outcomes of patients hospitalised 436
with COVID-19 in the UK show a signal towards an increased risk of day-28 mortality due to 437
HIV-positive status. The data for this study were collected during the peak of the UK 438
COVID-19 epidemic and the analysis contains a significant proportion of missing data, 439
including a high number of patients with missing HIV status, who were excluded from the 440
analysis. As the pandemic continues to spread in areas of increased HIV prevalence, our 441
observations highlight the importance of recording the HIV status of people hospitalised with 442
COVID-19 to ensure appropriate management during hospitalisation and gather further data 443
to improve our understanding of the reciprocal interactions between SARS-CoV-2 and HIV. 444
445
Despite effective ART and normalised CD4 cell counts, a subset of PWH continue to 446
experience immune activation, inflammation and a pro-coagulatory state,7 which may be 447
postulated to modulate the risk of COVID-19 related morbidity and mortality.28,29 One 448
determinant of such persistent immune dysfunction is the degree of immunosuppression 449
experienced prior to the start of ART, defined by a low nadir CD4 cell count and inverted 450
CD4:CD8 ratio. In the UK, 43% of people newly diagnosed with HIV in 2018 had a CD4 451
count <350 cells/mm3, a threshold indicative of significant immunosuppression.26 452
Furthermore, current guidelines about starting ART immediately at the time to diagnosis 453
were implemented relatively recently, whereas in the past ART initiation was deferred until 454
the CD4 count had declined below thresholds of initially 200, then 350 and subsequently 500 455
cells cells/mm3.6,30 Thus, many PWH in the UK and worldwide will have experienced years 456
of uncontrolled HIV replication prior to commencing treatment, and may have experienced 457
earlier regimens of suboptimal efficacy, with lasting effects on immune function. Planned 458
linkage of the hospital dataset with the HIV clinic records will be required to clarify the role 459
of ART history, current and nadir CD4 cell count, plasma HIV-1 RNA load and previous 460
history of HIV-related disease on the outcomes observed in this study. Meanwhile, emphasis 461
for PWH should be placed on early HIV diagnosis, prompt ART initiation, and optimised 462
screening for and control of comorbidities including obesity and diabetes. 463
464
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
abdominal pain. bBased on the onset of symptoms relative to the date of admission, COVID-19 471
acquisition was classed as community (<3 days), indeterminate (3-7 days), probable hospital (8-14 days), 472
and definite hospital (>14 days). cDefined as HR >100 beats/min. dDefined as RR >20 breaths/min. 473 eDefined as SpO2 <94% on air; proportions with SpO2 <94% or on oxygen therapy at presentation were 474
30/108 (27.8%) and 32/105 (30.5%) respectively in the HIV-positive group, and 14,463/45,123 (32.1%) 475
and 14,203/44,093 (32.2%) in the HIV-negative group. Abbreviations: IQR, Interquartile range; HR, 476
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
Table 4. Cox proportional hazards model of the association between HIV status and day-28 487
mortality 488
HIV-positive versus HIV-negative
Hazard ratio
95% CI P-value
Unadjusted 0.74 0.50-1.09 0.12 Adjusted for sex 0.71 0.48-1.05 0.08 Adjusted for ethnicity 0.77 0.52-1.15 0.21 Adjusted for age 1.39 0.94-2.09 0.10 Adjusted for age and sex 1.39 0.93-2.08 0.11 Adjusted for sex, ethnicity, age, baseline date, and indeterminate/probable hospital acquisition of COVID-19
1.49 0.99-2.25 0.06
Adjusted for sex, ethnicity, age, baseline date, indeterminate/probable hospital acquisition of COVID-19, and 10 comorbiditiesa
1.49 0.99-2.26 0.06
Adjusted for sex, ethnicity, age, baseline date, indeterminate/probable hospital acquisition of COVID-19, 10 comorbiditiesa, and hypoxia at presentationb
1.62 1.06-2.46 0.02
Adjusted for sex, ethnicity, age, baseline date, indeterminate/probable hospital acquisition of COVID-19, 10 comorbiditiesa and hypoxia/ receiving oxygen at presentationb
1.63 1.07- 2.48 0.02
aThe model adjusted for the following comorbidities: chronic cardiac disease, chronic 489
dementia. liver disease, malignancy, and chronic haematological disease. bHypoxia was 491
defined as SpO2 <94% on air; a record of hypoxia or receiving oxygen at presentation were 492
used as an indicator of disease severity. 493
494
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
Figure 2. Kernel density plot of age distribution of study participants stratified by HIV status. 508
Figure 3. Interventions during hospitalisation by HIV status. 509
Figure 4. Kaplan Meier survival graphs, stratified by HIV status, sex and age group. P values 510
represent log-rank tests. Plots D, E and F include only individuals from age groups <50 years, 511
50-59 years and 60-79 years. 512
513
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
Missing date of admission or symptom onset, n=2,694
Admission or symptom onset date >4th June 2020, n= 303
Missing HIV status, n=2,742
Unable to confirm HIV-positive status, n=63
Included
n=47,539
HIV-positive people, n=115
HIV-negative people, n=47,424
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
Figure 2. Kernel density plot of age distribution of study participants stratified by HIV status. 517
518
519
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
Figure 4. Kaplan Meier survival graphs, stratified by HIV status, sex and age group. P values 522
represent log-rank tests. Plots D, E and F include only individuals from age groups <50 years, 523
50-59 years and 60-79 years. 524
525
526
527
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
Annette Lake, Claire Petersen, and Scott Mullaney. 557
558
FINANCIAL SUPPORT 559
AS is supported by a National Institute of Health Research (NIHR) Academic Clinical 560
Lectureship at the University of Liverpool. LT is supported by the Wellcome Trust (grant 561
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
number 205228/Z/16/Z). The work is supported by grants from: the NIHR [award CO-CIN-562
01]; the Medical Research Council [grant MC_PC_19059]; the NIHR Health Protection 563
Research Units (HPRU) in i) Emerging and Zoonotic Infections (NIHR200907) at University 564
of Liverpool in partnership with Public Health England (PHE) and in collaboration with the 565
Liverpool School of Tropical Medicine and the University of Oxford, and ii) Blood Borne 566
and Sexually Transmitted Infections at University College London UCL in partnership with 567
PHE and in collaboration with the London School of Hygiene and Tropical Medicine; the 568
Wellcome Trust and the Department for International Development [215091/Z/18/Z], and the 569
Bill and Melinda Gates Foundation [OPP1209135]. The Liverpool Experimental Cancer 570
Medicine Centre provided infrastructure support for this research (Grant Reference: 571
C18616/A25153). The views expressed are those of the authors and not necessarily those of 572
the DHSC, DID, NIHR, MRC, Wellcome Trust or PHE. 573
574
DISCLOSURES 575
AMG: Personal fees from Roche Pharma Research & Early Development (pRED), consulting 576
honoraria from Gilead, Janssen, and ViiV Healthcare, and research funding from Roche 577
pRED, Gilead, Janssen and ViiV Heathcare, outside of the work presented in this article. GV: 578
research funding from ViiV Healthcare outside of the work presented in this article. CAS: 579
personal fees from Gilead Sciences and ViiV Healthcare for participation in Data Safety and 580
Monitoring Boards, membership of Advisory Boards and for preparation of educational 581
materials, outside of the work presented in this article. MGS: grants from DHSC NIHR UK, 582
MRC UK, and HPRU in Emerging and Zoonotic Infections during the conduct of the study; 583
other from Integrum Scientific LLC (Greensboro, NC, US), outside the submitted work. The 584
remaining authors declare no competing interests, no financial relationships with any 585
organisations that might have an interest in the submitted work in the previous three years, 586
and no other relationships or activities that could appear to have influenced the submitted 587
work. 588
589
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
deprivation, and hospitalization for COVID-19 in English participants of a national 620
biobank. Int J Equity Health. 2020;19:114. 621
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
Griffiths. Data Analysts: Lisa Norman, Riinu Pius, Tom M Drake, Cameron J Fairfield, 679
Stephen Knight, Kenneth A Mclean, Derek Murphy, Catherine A Shaw. Data and 680
Information System Manager: Jo Dalton, Michelle Girvan, Egle Saviciute, Stephanie Roberts 681
Janet Harrison, Laura Marsh, Marie Connor. Data integration and presentation: Gary 682
Leeming, Andrew Law, Ross Hendry. Material Management: William Greenhalf, Victoria 683
Shaw, Sarah McDonald. Outbreak Laboratory Volunteers: Katie A. Ahmed, Jane A 684
Armstrong, Milton Ashworth, Innocent G Asiimwe, Siddharth Bakshi, Samantha L Barlow, 685
Laura Booth, Benjamin Brennan, Katie Bullock, Benjamin WA Catterall, Jordan J Clark, 686
Emily A Clarke, Sarah Cole, Louise Cooper, Helen Cox, Christopher Davis, Oslem 687
Dincarslan, Chris Dunn, Philip Dyer, Angela Elliott, Anthony Evans, Lewis WS Fisher, Terry 688
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint
Anil Sharma, Anna Shawcross, Jeremy Sizer, Richard Smith, Catherine Snelson, Nick 723
Spittle, Nikki Staines , Tom Stambach, Richard Stewart, Pradeep Subudhi, Tamas Szakmany, 724
Kate Tatham, Jo Thomas, Chris Thompson, Robert Thompson, Ascanio Tridente, Darell 725
Tupper - Carey, Mary Twagira, Andrew Ustianowski, Nick Vallotton, Lisa Vincent-Smith, 726
Shico Visuvanathan , Alan Vuylsteke, Sam Waddy, Rachel Wake, Andrew Walden, Ingeborg 727
Welters, Tony Whitehouse, Paul Whittaker, Ashley Whittington, Meme Wijesinghe, Martin 728
Williams, Lawrence Wilson, Sarah Wilson, Stephen Winchester, Martin Wiselka, Adam 729
Wolverson, Daniel G Wooton, Andrew Workman, Bryan Yates, Peter Young. 730
All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprintthis version posted August 11, 2020. ; https://doi.org/10.1101/2020.08.07.20170449doi: medRxiv preprint